The successful implementation of evidence-based infection prevention practices (such as "maximum barrier precautions for central line insertion") is known to significantly reduce hospital-acquired infections (HAIs), mortality, and costs. However, while some hospitals have successfully implemented these national practice standards, others have failed in their attempts. The theoretical literature on organizational change has underscored the importance of "communication network structures," i.e., the direction and frequency of organizational communication, in enabling tacit knowledge exchange, learning, and improvement. Preliminary health services research in this area suggests that "top-down" communication network structures, where changes to work practices are initiated by those with authority (like administrators), may be most effective for tacit knowledge exchange, learning, and improvement in healthcare organizations. By contrast, recent anecdotal evidence from hospital infection prevention success stories, suggests that "peer-to-peer" communication network structures, where professionals from different subgroups (like physicians and nurses) directly communicate with each other on practice changes, with minimal interference from PIity, may be most effective for improvement on evidence-based practices. These inconsistencies suggest a lack of systematic evidence on which communication network structures are more effective for infection prevention, i.e., which are associated with higher compliance on evidence-based practices and lower HAI rates. Within the context of HAIs, this pilot/feasibility study focuses on central line blood stream infections (CLBSIs). The first aim of this study is to develop methods for measuring the "communication network structure," "content of communication," and "outcomes" related to evidence-based CLBSI prevention practices at the unit level. The "communication network structure" refers to the direction and frequency of communication on CLBSI prevention practices across various professional subgroups and hierarchical levels, including medical faculty, nurses, residents, students, unit managers and hospital administrators. The "content of communication" refers to the type of knowledge (i.e., "tacit" vs. "explicit" knowledge) exchanged on CLBSI practices. "Outcomes" include compliance with CLBSI prevention practices and hospital-acquired CLBSI rates at the unit level. The second aim of the study is to conduct an initial test of the hypothesis that "top-down communication network structures are associated with better infection prevention outcomes." The setting will be two intensive care units in an academic medical center. Data on "communication network structure" and "content of communication" will be collected weekly using "communication logs" completed by participants in each subgroup/level. Data on unit "outcomes" will be collected weekly through medical record review. All data will be collected before and after an "organizational pledge" to improve performance on CLBSI prevention practices. In all, data will be collected over 52 weeks in two units, resulting in 104 unit-week observations. Analysis will include content analysis of types of knowledge exchanged on CLBSI practices;network analysis of communication network structures (alongside unit outcomes);and regression analysis of the relationship between communication network structures and outcomes. PUBLIC HEALTH RELEVANCE: According to the Agency for Healthcare Research and Quality (AHRQ, 2009a), nearly 2 million patients develop hospital-acquired infections (HAIs), which contribute to 99,000 deaths each year and $28 billion to $33 billion in health care costs. The successful implementation of evidence-based infection prevention practices is known to significantly reduce HAIs. This study has potential to make substantive contributions to public health by enabling the successful implementation of evidence-based practice standards in healthcare organizations. In addition to the public at large, study results would be directly beneficial to a variety of stakeholders, including healthcare managers &professionals, accreditation agencies, policy makers, and health service researchers.